Dielectrophoretic Forces on the Nanoscale - American Chemical Society

as the particle. The Boundary Element Dielectrophoretic Force (BEDF) method involves constructing a ... BEDF and PDA methods for a 1 Е polarizable sp...
0 downloads 0 Views 737KB Size
Download PDF
10744

Langmuir 2004, 20, 10744-10750

Dielectrophoretic Forces on the Nanoscale C. M. Schaldach,† William L. Bourcier,† Phillip H. Paul,‡ and W. D. Wilson*,† Lawrence Livermore National Laboratory, Livermore, California 94550, and Eksigent Technologies, Livermore, California 94550 Received April 7, 2004. In Final Form: August 20, 2004 We have developed a method of calculation of the dielectrophoretic force on a nanoparticle in a fluid environment where variations in the electric field and electric field gradients are on the same nanoscale as the particle. The Boundary Element Dielectrophoretic Force (BEDF) method involves constructing a solvent-accessible or molecular surface surrounding the particle, calculating the normal component of the electric field at the surface boundary elements, and then solving a system of linear equations for the induced surface polarization charge on each element. Different surface elements of the molecule may experience quite different polarizing electric fields, unlike the situation in the point dipole approximation. A single 100-Å-radius ring test configuration is employed to facilitate comparison with the well-known point dipole approximation (PDA). We find remarkable agreement between the forces calculated by the BEDF and PDA methods for a 1 Å polarizable sphere. However, for larger particles, the differences between the methods become qualitative as well as quantitative; the character of the force changes from attractive at the origin of the ring for a 50-Å sphere, to repulsive for a 75-Å sphere. Equally dramatic differences are found in a more complex electrical environment involving two sets of 10 rings.

I. Introduction Dielectrophoresis is increasingly being employed to manipulate and separate molecules and particles including biological cells.1-7 Recently, Bennett, et al1 have driven dielectrophoretic phase separations of polystyrene beads and bacteria in microfluidic channels because of induced dipole-dipole interactions. Voldman, et al2 have investigated how dielectrophoresis-based quadrupole traps for single beads can hold against destabilizing fluid flows. Gascoyne, et al3,4 have separated cancer cells in blood using dielectrophoretic-driven levitation. On the basis of the same physics, optical tweezers have been applied to microspheres5 and cells6,7 and have even formed the basis of a Nobel Prize.8 Furthermore, recent developments in nanotechnology enable structures to be built which can create fields and field gradients on unprecedented length scales; the scale of the variations in the field inducing charge on a molecule may be the same as the scale of the molecule itself. Synthetic nanopores have been fabricated in inorganic materials for transporting DNA.9,10 Carbon nanotubes * Author to whom correspondence should be addressed. E-mail: [email protected]. Tel: 925-424-2497. † Lawrence Livermore National Laboratory. ‡ Eksigent Technologies. (1) Bennett, D. J.; Khusid, B.; James, C. D.; Galambos, P. C.; Okandan, M.; Jacqmin, D.; Acrivos, A. Appl. Phys. Lett. 2003, 83, 4866. (2) Voldman, J.; Braff, R. A.; Toner, M.; Gray, M. L.; Schmidt, M. A. Biophys. J. 2001, 80, 531. (3) Gascoyne, P. R. C.; Vykoukal, J. Electrophoresis 2002, 23, 1973. (4) Gascoyne, P. R. C.; Wang, X.-B.; Huang, Y.; Becker, F. F. IEEE Trans. Ind. Appl. 1997, 33, 670. (5) Wright, W. H.; Sonek, G. J.; Berns, M. W. Appl. Optics 1994, 33, 1735. (6) Fuhr, G.; Arnold, W. M.; Hagedorn, R.; Muller, T.; Benecke, W.; Wagner, B.; Zimmermann, U. Biochim. Biophys. Acta 1992, 1108, 215. (7) Schnelle, T.; Muller, T.; Hagedorn, R.; Voigt, A.; Fuhr, G. Biochim. Biophys. Acta 1999, 1428, 99. (8) Chu, S. for development of methods to cool and trap atoms with laser light; Nobel Prize in Physics, 1997; Optical tweezers reviewed in: Molloy, J. E.; Padgett, M. J. Contemp. Phys. 2002, 43, 241. (9) Meller, A.; Nivon, L.; Brandin, E.; Golovchenko, J.; Branton, D. Proc. Natl. Acad. Sci. 2001, 97, 1079. (10) Li, J.; Stein, D.; McMullan, C.; Branton, D.; Aziz, M. J.; Golovchenkok, J. Nature 2001, 412, 166.

have been aligned in a polymer film to demonstrate molecular transport through their cores.11,12 Dielectrophoresis has recently been employed to assemble nanowires in suspensions.13 Daiguji, et al,14 for example, have studied ion current flow through 30-nm silica channels which were gated at their midpoint. Using electroless deposition of Au in track-etched pores in polycarbonate films, Martin, et al15 have constructed membrane channels of the order of molecular dimensions in studies of permselectivity of membranes. Hybrid three-dimensional fluidic architectures have been designed which control nanofluidic transfer via surface-charged nanopores separating crossed microfluidic channels.16 Perhaps most notably, multilayer technology enables materials comprised of virtually any element to be constructed with control on atomic dimensions.17,18 In light of these developments in nanoengineering, it seems appropriate to investigate the fundamental assumptions underlying the analytical formulation of dielectrophoretic forces (DFs). We have previously presented a method for calculating molecular adsorption (H2S and HS- on Cu) utilizing boundary elements in a GouyChapman field.19 The method was shown to agree well with the analytic result for a dielectric sphere in a fluid environment.20 Here, we extend and apply the method to the calculation of DFs in situations where the scale of the variations in the field inducing charge on a molecule may (11) Hinds, B. J.; Chopra, N.; Rantell, T.; Andrews, R.; Gavalas, V.; Bachas, L. G. Science 2004, 303, 62. (12) Noy, A.; Bakajin, O. private communication. (13) Hermanson, K. D.; Lumsdon, S. O.; Williams, J. P.; Kaler, E. W.; Velev, O. D. Science 2001, 294, 1082. (14) Daiguji, H.; Yang, P.; Majumdar, A.; Nano Lett. 2004, 4, 137. (15) Martin, C. R.; Nishizawa, M.; Jirage, K.; Kang, M.; Lee, S. B. Adv. Mater. 2001, 13, 1351. (16) Kuo, T.-C.; Cannon, D. M., Jr.; Chen, Y.; Tulock, J. J.; Shannon, M. A.; Sweedler, J. V.; Bohn, P. W. Anal. Chem. 2003, 75, 1861. (17) Bajt, S.; Alameda, J. B.; Barbee, T. W.; Clift, W. M.; Folta, J. A.; Kaufmann, B.; Spiller, E. A. Opt. Eng. 2002, 41, 1797. (18) Spiller, E.; Baker, S. L.; Mirkarimi, P. B.; Sperry, V.; Gullikson, E. M.; Stearns, D. G. Appl. Opt. 2003, 42, 4049. (19) Wilson, W. D.; Schaldach, C. M. J. Colloid Interface Sci. 1998, 208, 546. (20) Jackson, J. D. Classical Electrodynamics, 3rd ed.; John Wiley and Sons: New York, 1999.

10.1021/la040059+ CCC: $27.50 © 2004 American Chemical Society Published on Web 10/30/2004

Dielectrophoretic Forces on the Nanoscale

Langmuir, Vol. 20, No. 24, 2004 10745

be the same as the scale of the molecule and whose molecular shape may not be spherical. The results are compared to the analytic (point dipole approximation, PDA) expressions for the DF.21-23 Sancho et al24 and Gheorghiu et al25 have recently developed similar methods involving boundary elements for studying shelled particles and cells. In Section II, we present our method of calculation. Section III contains a comparison between our method and analytic expressions for the DF when the electric field is due to a simply charged ring. Multiple-ring configurations are compared in Section IV. Section V contains a Summary and Conclusions. II. Method of Calculation

W)

{[I] - f[K]}[σ] ) f[E‚n]

by21-23

(1)

For a sphere of radius R having internal dielectric constant 2 in a dielectric medium of dielectric constant 1, the effective dipole moment is

(

)

 2 - 1 p ) 2πR31 E 2 + 21

(2)

where (2 - 1)/(2 + 21) is the well-known Clausius-Mossotti factor for a sphere.20 The DF in the PDA becomes

(

F ) 2πR31

)

2 - 1 ∇|E|2 2 + 21

(3)

As Pohl has pointed out,21 eq 3 is applicable to a “small” sphere (although not a point dipole since it has finite radius, R) in a field, E, which is assumed to be nonuniform enough to produce appreciably different charges on the positive and negative regions of the induced surface polarization charges but which nevertheless does not vary so strongly as to alter the size of the dipole throughout the sphere. The DF calculated using eq 3 does assume the molecule (“molecule” and “particle” are equivalent for our purposes here) to be a dipole: the magnitudes of the positiveand negative-induced polarization charges are equal. Equation 3 is recognized as being applicable to the force on a molecule when the dimensions of the variations in the electrostatic potentials and fields are small compared to molecular dimensions. An alternative expression for the force on a molecule is23

F)

∫σ(s)E(s)ds

(4)

where σ(s) is the induced surface charge density on an element, s, of the molecular surface and E(s) is the electric field at the surface element having elemental area ds. Equation 4 (a) is applicable to arbitrary molecular geometry and (b) takes account of the precise electric field and, hence, field gradient at every element of the molecule. A molecule whose surface elements experience electric fields whose gradients are not representable by the gradient at say, the centroid, may not be able to be represented by the PDA. Given the charges induced on the surface elements, the energy, W, required to bring the molecule from infinity (a position where the field or field gradient is zero) to its position in the nanostructure is given by (21) Pohl, H. A. Dielectrophoresis; Cambridge University Press: Cambridge, 1978. (22) Jones, T. B. Electromechanics of Particles; Cambridge University Press: Cambridge, 1995. (23) Jones, T. B. J. Electrostat. 1979, 6, 69. (24) Sancho, M.; Martinez, G.; Martin, C. J. Electrostat. 2003, 57, 143. (25) Gheorghiu, E. Ann. N. Y. Acad. Sci. 1999, 873, 262.

∫φ(s)σ(s)ds

(5)

where φ(s) is the electrostatic potential at the molecular surface element s. Our method of calculation, the Boundary Element Dielectrophoretic Force (BEDF) method, involves first constructing a molecular or solvent-accessible surface surrounding the molecule by a method we have previously described.19 This surface provides the interface between the dielectric media and the molecule; elements of the surface are assigned a unit normal and an elemental area. The electric field, E, created by the nanostructure provides a source of polarization of the molecule. The induced interfacial charge, σ, can be obtained from a straightforward consideration of the electrostatic boundary conditions and selfterms. This leads to a system of linear equations,26,19

The DF, F, on a dipole, p, in a nonuniform field, E, is given

F ) (p‚∇)E

1 2

(6)

where

Kik )

rik‚ni

∫ |r

3 ik|

dSk

(7)

rik is the vector distance between elements i and k on the molecular boundary, ni is the outward normal at boundary element i, dSk is the differential associated with the area of boundary element k, and E‚n is the column vector of normal components of the electric field. f is given by

f)

2 - 1 2π(1 + 2)

(8)

2 ) 1.0 here but is not restricted to this, and 1 is the dielectric constant of the solvent. 1 ) 78.5 for water here but, again, is not restricted to this. Solution of eq 6 by the usual methods of linear algebra provides the polarization charge, σ, created by E at each surface element. In this way, different surface elements of the molecule are allowed to experience quite different polarizing electric fields, unlike the situation in the PDA. The field gradient, variations in the field over the scale of the molecule, is properly taken into account.

III. Molecule in the Vicinity of a Ring A simple charged ring of radius R ) 100 Å provides an interesting test case. The axial electric field has variations over length scales of the order of its radius which serve to illustrate several important features. The potentials and fields can be calculated analytically20 and can also be obtained numerically by constructing the ring of small Debye-Hu¨ckel27 atoms (spheres) and performing a direct summation of their individual contributions.28 We employed the numerical method here to facilitate our investigation into more-complicated structures to be described below. In these demonstration calculations, each atomic element of the ring was given a charge of 0.1 electrons; the magnitude of the fields and forces in this test case are small but obviously scale with the charge. In Figure 1, we have plotted the magnitude of the electric field, |E|, due to the charged ring of radius R ) 100 Å (the ring is in the x-y plane, centered at (x,y,z) ) (0,0,0)) as a function of the distance, z, (a) along its axis, x ) y ) 0.0 (solid line, filled circles) and also (b) along an axis at x ) 75.0 Å, y ) 0 (solid line alone). It is interesting to note that |E| along the x ) 0 axis has peaks, maxima, at zm ) ( (26) Zauhar, R. J.; Morgan, R. S. J. Mol. Biol. 1985, 186, 815. See also: J. Comput. Chem. 1990, 11, 603; J. Comput. Chem. 1988, 9, 171; J. Comput. Chem. 1991, 12, 575. (27) Debye, P.; Hu¨ckel, E. Physik Z. 1923, 24, 185. (28) Schaldach, C. M.; Bourcier, W. L.; Paul, P. H.; Wilson, W. D. J. Colloid Interface Sci. 2004, 275, 601.

10746

Langmuir, Vol. 20, No. 24, 2004

Schaldach et al.

Figure 1. Magnitude of the electric field, |E|, due to a charged ring of radius R ) 100 Å as a function of the distance, z, (a) along its axis, x ) y ) 0.0 (solid line, filled circles) and also (b) along an axis at x ) 75.0 Å, y ) 0 (solid line alone). Note that |E| along the x ) 0 axis has maxima at zm ) ((x2/2)R (∼70.7 Å); along the x ) 75.0 Å axis, a single maximum in |E| is seen at z ) 0.0. This shift in the peak of the magnitude of the field is caused by the increase in the off-axis x component at x ) 75.0 Å (short-dashed line). The z component of the electric field along the x ) 75.0 Å axis (long-dashed line), changes sign at z ) 0 and has its own extrema at ( 25.4 Å ((48 Bohr); nevertheless, the x component, peaked at z ) 0.0, dominates the field magnitude off-axis (x ) 75.0 Å).

(x2/2)R (∼70.7 Å) due to the symmetry-induced dominance of the z component of the field. On the other hand, along the x ) 75.0 Å axis, a single maximum in |E| is seen at z ) 0.0. This shift in the peak of the magnitude of the field is caused by the increase in the off-axis x component at x ) 75.0 Å, also plotted in Figure 1 (short-dashed line). Note that the x component also exhibits small peaks at z ) 84.1 Å (150 Bohr). (The y component of the electric field is zero in both cases because of the symmetry.) The z component of the electric field along the x ) 75.0 Å axis, also shown in Figure 1 (long-dashed line), changes sign at z ) 0 and has its own extrema at (25.4 Å ((48 Bohr); nevertheless, the x component, peaked at z ) 0.0, dominates the field magnitude off-axis (x ) 75.0 Å). Next, the DF on a sphere of radius Rm ) 1 Å was calculated as a function of its axial distance, z, along the ring axis (x ) y ) 0.0) using both the PDA (eq 3 above) and the BEDF method (eq 4 above). The results of this comparison are shown in Figure 2. It is first of all to be noted that, for this small molecule, these very different methods of calculation are in remarkable agreement with each other. The DF is negative for small z > 0 (molecule is being attracted backward toward the ring center) and positive for small z < 0 (molecule is being attracted toward the ring center), indicating that the molecule is trapped at the ring center. As we will discuss in greater detail later, in the region -zm < z < zm, the induced polarization charge on the surface elements of the sphere nearest the ring center is found by direct solution to eq 6 to be positive, while those elements furthest from the ring are determined to be negatively charged. The electric field in this region is growing in magnitude (see Figure 1); the force on the negative charge therefore exceeds that on the positive charge, making the total force negative, as indicated in Figure 2. When the molecule is along the z axis at z > zm, the sign of the induced polarization charges is as for z < zm, but now the magnitude of the electric field is diminishing. Consequently, the positive surface charge interacts with a larger electric field, and the total force

Figure 2. Dielectrophoretic force (DF) on a sphere of radius Rm ) 1 Å as a function of its axial distance, z, along the ring axis (x ) y ) 0.0) using both the PDA (eq 3) and the BEDF method (eq 4). Note that, for this small molecule, these very different methods of calculation are in remarkable agreement with each other. The dielectrophoretic energy calculated from the induced charges and the electrostatic potential (eq 5) (solid line, open circles) exhibits a minimum in the energy where the force has a negative slope as it goes through zero, but an energy maximum occurs when the force goes through zero with positive slope.

on the molecule becomes positive. Note from Figure 2 that this occurs at z ) zm where the peak in the magnitude of the electric field, |E|, is seen to occur (see Figure 1). Similar effects occur at z < -zm. Whereas the region near the ring origin is attractive, the region near the extrema in the electric field is repulsive. The dielectrophoretic energy calculated from the induced charges and the electrostatic potential (eq 5 above), also plotted in Figure 2 (solid line,

Dielectrophoretic Forces on the Nanoscale

Langmuir, Vol. 20, No. 24, 2004 10747

Figure 3. Dielectrophoretic force on spheres of radius 50 Å and 75 Å as a function of distance, z, along the ring axis calculated in the PDA and by the BEDF method. The two methods give similar characteristics, albeit different magnitudes, for the 50-Å sphere. For the 75-Å sphere, however, the slope of the force vs distance curve changes sign. This is a quantitative and qualitative change in the force from attractive for the 50-Å molecule to repulsive for the 75-Å molecule. The PDA predicts both molecules to be attracted toward the center. In Figure 3b, we expand the inset region noted in Figure 3a for 50-Å and 75-Å spheres and include the dielectrophoretic force for a 65-Å sphere, which shows the transition in the slope of the force taking place as a function of sphere radius.

open circles), reflects these effects: For example, a minimum in the energy is seen to occur where the force has a negative slope as it goes through zero, but an energy maximum occurs when the force goes through zero with positive slope. In Figure 3a, we have plotted the DF on spheres of radius 50 Å and 75 Å as a function of distance, z, along the ring axis (x ) y ) 0), again calculated in the PDA and by the BEDF method. The two methods give similar characteristics, albeit different magnitudes, for the 50-Å sphere: positive force for small z < 0 and negative force for small z > 0 as found for the 1-Å sphere in Figure 2. As for the 1-Å sphere, the PDA predicts higher-magnitude forces because of the x component of the off-axis electric field at the molecular surface elements. At x ) 75 Å, the surface elements of the sphere experience dramatically different electric fields (even a shift in the peaks of |E|, see Figure 1) which dramatically change the character of the DF. We see in Figure 3a that the slope of the force vs distance curve changes sign. This is a quantitative and qualitative change in the force from attractive for the 50-Å molecule to repulsive for the 75-Å molecule. The PDA predicts both molecules to be attracted toward the center. In addition, the PDA predicts repulsion at the peaks in the field, i.e., the slope of the DF vs distance curve is

positive as the force goes through zero, while the BEDF method exhibits no such repulsive behavior at the peaks. In Figure 3b, we expand the inset region noted in Figure 3a for 50-Å and 75-Å spheres and include the DF for a 65-Å sphere. This intermediate-sized sphere shows a weak attraction, reduced by an order of magnitude from the force on the 50-Å sphere and shows the transition in the slope of the force taking place as a function of sphere radius. It is interesting to note from Figure 3b that the DF on a 50-Å sphere moved along z at the x ) 30 Å off-axis position exhibits similar behavior to the 65-Å sphere moved onaxis. As the off-center sphere moves far from the ring center (not shown), it follows the behavior for the 50-Å on-axis sphere. As we have alluded to above, a better understanding of these effects involves the induced surface polarization charges (more precisely, the normal component of the electric field at molecular surface elements, see eq 6.) In Figure 4, we have plotted the positive and negative surface polarization charges on the elements of the 1-Å and 75-Å spheres as a function of the position of the centroid of each molecule. (These charges are obviously separated from each other.) For the 1-Å sphere, we find two extrema each for the positive and negative charges, these extrema coinciding with the maxima in the magnitude of the electric

10748

Langmuir, Vol. 20, No. 24, 2004

Figure 4. Positive and negative surface polarization charges on the elements of the 1-Å and 75-Å spheres as a function of the position of the centroid of each molecule. For the 1-Å sphere, we find two extrema each for the positive and negative charges, these extrema coinciding with the maxima in the magnitude of the electric field, zm, for the 1-Å sphere shown in Figure 1. For the 75-Å sphere, the positive and negative induced charges have a markedly different character: a single extremum at the ring origin (z ) 0) is found for each.

field, zm, for the 1-Å sphere shown in Figure 1. For the 75-Å sphere, the positive and negative induced charges have a markedly different character: a single extremum at the ring origin (z ) 0) is found for each. For the small sphere, the negative charge induced on the leading edge of the molecule as it moves toward positive z values interacts with the large electric field at those leading elements, resulting in a negative force. The positive charge induced by the x component of the field on the small sphere multiplied by the z component of the field at those elements is insufficient to overcome this negative force. On the other hand, in the case of the 75-Å sphere, there is a large increase in positive charge on elements of the sphere near the origin (actually, from the large normal component, E‚n, see eq 6 and Figure 1), and now this increase in positive charge, even though multiplied by a smaller field, is sufficient to overcome the larger field at the leading negative elements of the molecule. The character of the force changes from attractive to repulsive. An examination of the details of the induced charges on the individual surface elements of each sphere in the ring environment provides further understanding of the change in character of the force. Specifically, with the 75-Å sphere at the origin, surface elements nearest to the origin (those having |z| < ∼24 Å) have positive charges induced; negative charges are induced on those elements furthest from the origin (|z| > ∼24 Å). The x and y components of the electric field are responsible for this large surface charge being induced. (From Figure 1 (solid line alone), note the large magnitude of the electric field along the x ) 75 Å axis.) If the field were entirely along the z axis, the magnitude of these induced charges on the 75-Å sphere would be reduced by an order of magnitude. In Figure 5, then, we have plotted the total induced charges for the 1-Å and 75-Å spheres when the field is (fictitiously) entirely along the z axis. Here, we find the nature of the induced surface charge on the small and large spheres to be quite similar; elements of the 75-Å sphere near the origin of the ring no longer become highly charged. These calculations provide direct evidence of the influence of the perpendicular

Schaldach et al.

Figure 5. Induced charges on each of the 1-Å and 75-Å spheres as a function of the position of their centroids along the ring axis when the field is (fictitiously) entirely along the z-axis. In contrast to Figure 4, here we find the nature of the induced surface charge on the small and large spheres to be quite similar; elements of the 75-Å sphere near the origin of the ring no longer become highly charged. These calculations provide direct evidence of the influence of the perpendicular component of the field on the induced charges and, hence, the character of the dielectrophoretic force.

component of the field on the induced charges and, hence, the character of the DF. The PDA assumes the molecule to be spherical (Clausius-Mossotti); the BEDF method presented here allows us to investigate shape effects. To this end, using the method described above, we constructed a molecular surface surrounding a 10 atom × 10 atom planar molecule (“face-centered-cubic-like” geometry, radius of each atom ) 1.2 Å) and then calculated the DF on that planar molecule in the same ring environment as above. In Figure 6, we present the results of these calculations when the plane of the molecule is the same as the ring (labeled “parallel” in Figure 6) and also when the molecule is rotated 90° about the x axis so its normal is perpendicular to the plane of the ring (“perpendicular” in Figure 6). Both orientations give rise to attractions to the origin and repulsions at the maxima of the electric field (zm in Figure 1), the force changing sign appropriately at these critical points. The parallel orientation results in forces which can be a factor of 2-3 greater than the perpendicular configuration. From Figure 6, we find the binding energy (difference in energy between the energy maxima and minima) for the parallel configuration is nearly twice the binding of the perpendicular geometry. These calculations indicate that shape effects must be considered when calculating DFs on this scale. IV. Molecule in a Complex Ring Environment As discussed above, nanotechnology has enabled interesting and technologically relevant geometrical configurations to be produced. Notable among these are multilayers or nanolaminates17,18 where atomic-scale layers of different materials can be produced adjacent to one another with single-atom interfaces between them. Optimizing a nanoscale geometry is beyond the scope of this work: we calculated the electric fields for a set of rings which is illustrative of the enabling power of the technology while providing further examples of the need

Dielectrophoretic Forces on the Nanoscale

Figure 6. Shape effects: DF on a planar molecule in the same ring environment as in Figure 1 when the plane of the molecule is the same as the ring (labeled “parallel”) and also when the molecule is rotated 90° about the x axis so its normal is perpendicular to the plane of the ring (“perpendicular”). The parallel orientation results in forces which can be a factor of 2-3 greater than the perpendicular configuration. The binding energy (difference in energy between the energy maxima and minima) for the parallel configuration is nearly twice the binding of the perpendicular geometry. These calculations indicate that shape effects must be considered when calculating DFs on this scale.

to calculate DFs using a molecular theory appropriate to these nanoscale configurations. When multiple rings are employed, the fields become more interesting. N rings of the same charge placed next to each other result in both an increase in the magnitude of the resulting field and a decrease in the position of the extrema relative to the plane of the end ring. An oppositely charged set of N rings could be configured along the same

Langmuir, Vol. 20, No. 24, 2004 10749

axis as the first at a distance chosen to optimize the magnitude of the field between the sets of rings. A single set of 10 coaxial rings, axis along z, radius ) 100 Å, comprised of atoms, each having a charge of +0.1 electrons as for the single ring above, placed next to each other in a row, will produce the electric field profile shown in Figure 7. The set, or layer, has a charge density of ∼3.4 µC/cm2. The “leftmost” ring is placed (in the x-y plane) at z ) 0; the electric field is zero in the middle of the 10-ring configuration (at 19 Bohr ) 10.1 Å) and directed positively along the z axis. A second 10-ring configuration, this time negatively charged, coaxial with the first (see Figure 7) will also produce an electric field directed positively along the z axis. By spacing the 10-ring sets so the extrema in the z component of their electric fields coincide (separation ) 123.3 Å (233 Bohr)), we optimize the field at 154 Bohr (81.5 Å), the midpoint between the rings (see Figure 7). As can be seen in Figure 7, this 2 ring × 10 ring configuration also gives rise to zeroes in the field at -15.3 Å (-29 Bohr) and 177.8 Å (336 Bohr), as well as secondary peaks at -73.6 Å (-139 Bohr) and 236.0 Å (446 Bohr). The off-axis electric field for this 2 ring × 10 ring configuration is dramatically different from that along the axis. In Figure 7, we have also plotted the x component and magnitude of the electric field for the 2 ring × 10 ring configuration along the z axis at x ) 75 Å. The large x component in the middle of the rings is responsible for the double-peaked character of the magnitude. In Figure 8, we show the results of our DF calculations for a 75-Å and a 90-Å sphere in the 2 ring × 10 ring environment described above using the BEDF method and, for comparison, the PDA method for the 75-Å sphere. (The PDA method for the 90-Å sphere shows the same behavior as that of the 75 Å, with its magnitude adjusted for the radius.) The PDA predicts an attraction (negative slope of DF vs z as it goes through zero) where the electric field is zero (-29 Bohr and 336 Bohr) and a repulsion

Figure 7. Electric field profile due to single set of 10 coaxial rings (closed circles), axis along z, radius ) 100 Å, comprised of atoms each having a charge of +0.1 electrons. The “leftmost” ring is placed (in the x-y plane) at z ) 0; the electric field is zero in the middle of the 10-ring configuration. A second 10-ring configuration, this time negatively charged, coaxial with the first (closed squares), is spaced so the extrema in the z component of their electric fields coincide; we optimize the field at 154 Bohr (81.5 Å), the midpoint between the rings. The off-axis electric field for this 2 ring × 10 ring configuration is dramatically different from that along the axis (dashed and closed lines). The large x component in the middle of the rings is responsible for the double-peaked character of the magnitude of the field at x ) 75 Å.

10750

Langmuir, Vol. 20, No. 24, 2004

Schaldach et al.

the force on a 90-Å sphere: The DF on a 90-Å sphere in the field maximum is attractive according to the BEDF method. The PDA is probing the field and field gradient along the axis of the 2 ring × 10 ring configuration, while the molecular surface elements (BEDF) at which the charge is being induced are probing quite different regions because of the scale of the field variations (see Figure 7). V. Summary and Conclusions

Figure 8. DF calculations for a 75-Å and a 90-Å sphere in a 2 ring × 10 ring environment (see Figure 7) using the BEDF method and, for comparison, the PDA method for the 75-Å sphere. Strikingly, at the maximum in the |E| field for the 2 ring × 10 ring, (see x ) 0.0 |E| field (z ) 154 Bohr) in Figure 7), the BEDF and PDA methods applied to the 75-Å sphere agree as to the repulsive character of the force. However, the methods produce opposite characters for the force on a 90-Å sphere: The DF on a 90-Å sphere in the field maximum is attractive according to the BEDF method. The PDA is probing the field and field gradient along the axis of the 2 ring × 10 ring configuration, while the molecular surface elements (BEDF) at which the charge is being induced are probing quite different regions because of the scale of the field variations (see Figure 7).

(positive slope of DF vs z as it goes through zero) where the secondary peaks in |E| occur (see Figure 7). The BEDF method predicts no attractions or repulsions in these regions. Strikingly, at the maximum in the |E| field for the 2 ring × 10 ring, (see x ) 0.0 |E| field (z ) 154 Bohr) in Figure 7), the BEDF and PDA methods applied to the 75-Å sphere agree as to the repulsive character of the force (positive slope of DF vs z as it goes through zero). However, the methods produce opposite characters for

With the advent of modern engineered nanostructures, it is vital to employ a method of calculation of the DFs on nearby molecules which is commensurate with the scale of the structure. The PDA was never intended to be employed where variations in the electric field occur on the same scale as that of the molecule or particle.21-23 We find remarkable agreement between the forces calculated by our BEDF method and those obtained in the PDA for small molecules. For larger molecules, however, we find quantitative and qualitative differences between the methods. We have presented results for a simple theoretical example in this work, that of a ring, which clearly show the pitfalls in employing the PDA: An excess of charge induced on the molecule can lead to a change in the character of the force in various regions of the ring, attractive vs repulsive, in direct disagreement with the PDA. Furthermore, shape effects are naturally included in the BEDF method and, for a simple example of a planar 10 atom × 10 atom molecule, found to introduce a factor of ∼2-3 difference in the force, depending upon orientation of the molecule. Beyond the simple ring example, we considered a possible nanoengineered structure motivated by the technology of multilayers, that of two oppositely charged layers of 10 rings each. Here, the forces are large enough to be of technological relevance. Again, we find dramatic differences in the character of the forces between our BEDF and the PDA methods. While attractions and repulsions are found in regions where the electric field changes sign using the PDA, no such forces are obtained for the BEDF. Because of the subtleties involved in predicting the induced surface charge on elements of a molecule, it is difficult to know beforehand where the PDA might break down. This necessitates the use of the theory appropriate to the scale of the problem. LA040059+